On board health monitoring and control of gas turbine engines
The reliability and performance of power generation systems, such the gas turbine engines, are key issues on which modern engine manufacturers and aircraft companies always focus. Studying the behaviour of an engine during its lifetime has led to the development of numerous models and tools for engine condition monitoring and adaptive control. Although these tools offer substantial cost savings to the industries and airline companies, most of them have been proven inefficient in their use. This is mainly due to the fact that most of the currently available diagnostic and control systems resulted from simple linear thermodynamical stationary models. Moreover, the simulations usually run on ground computers, which makes the diagnosis, fault identification and adaptive control difficult to implement on board. Urged by this, this project developed a set of tools and methodologies for a new generation of controllers and systems for increasing both reliability and performances of engines. More specifically, a new complex non-linear dynamical model running on small on-board computers has been generated that improves the accuracy of diagnosis and monitoring system. The project extensively analysed four different diagnosis and fault identification techniques and employed mathematical methods grounded on neural networks. Hence, it resulted in a well-established tool for performing an automatic diagnosis of a reference turbine engine. These methodologies coupled with the non-linear model resulted in an enhanced adaptive fault tolerance control that leads to maximisation and optimisation of engine performance, fault tolerance and lifetime of components. The project's results contribute to the timely, reliable and accurate diagnosis and control of the engine's performance. Furthermore, this project specified the necessary common basis for implementing the future generation of controls and systems with increased safety, reliability and performances of power generation systems.
